scieee Science in your language
[en] (orig)

DC Motor Simulation with ARM Based Hardware in the Loop

Abstract

Hardware in the loop testing is increasingly important product testing for cutting down time to market. In my previous article I identified the parameters of a DC motor control system using system identification. An implementation of the resulting transfer function was developed on an ARM based hardware in the loop system and has been verified against the original system. The results show that the hardware in the loop system produces responses within tolerance to the stimulus signals and can be used for testing of controlsystems.

Read accessible full text

DC Motor Simulation with ARM Based Hardware in the Loop

Author: Csikós, Sándor; Bálint, Ádám
Publisher: DUPress
Year: 2016
Source: https://dea.lib.unideb.hu/bitstreams/ed16b646-0587-4221-aeaf-318d4ef1c254/download
Recen Inno a ions in Mecha onics (RIiM) Vol. 3. (2016). No. 1-2.
DOI: 10.17667/ iim.2016.1-2/15.
DC Mo o Simula ion wi h ARM Based
Ha dwa e in he Loop
Sándo Csikós
Uni e si y o Szeged Facul y o Enginee ing
TechnicalIns i u e
Szeged, Hunga y
[email protected]
Ádám Bálin
Uni e si y o Szeged Facul y o Enginee ing
Technical Ins i u e
Szeged, Hunga y
Abs ac —Ha dwa e in he loop es ing is inc easingly
impo an p oduc es ing o cu ing down ime o ma ke . In
my p e ious a icle I iden i ied he pa ame e s o a DC mo o
con ol sys em using sys em iden i ica ion. An implemen a ion o
he esul ing ans e unc ion was de eloped on an ARM based
ha dwa e in he loop sys em and has been e i ied agains he
o iginal sys em. The esul s show ha he ha dwa e in he loop
sys em p oduces esponses wi hin ole ance o he s imulus
signals and can be used o es ing o con olsys ems.
Keywo ds—Ha dwa e in he loop, LabVIEW, ARM
mic ocon olle
I. INTRODUCTION
Model based design is a me hodology used in designing
and es ing con olle so wa e implemen ed in PLCs o
mic ocon olle s. The h ee phases ha phases o model based
design a e model in he loop (MIL) whe e only he
ma hema ical ep esen a ions o he plan and con ol sys em
a e used, so wa e in he loop (SIL) whe e a simula ed
implemen a ion o he con olle con ols a simula ion o he
plan and ha dwa e in he loop (HIL) whe e a eal con olle
con ols a eal- ime simula ion o a plan . As he las
simula ion s age o p oduc de elopmen ha dwa e in he loop
simula ion is used o cu down es ing ime, ensu e he
co ec ness o he implemen ed con ol algo i hm and
dec ease es ing cos s. As such i is impo an o make he
simula ing de ice be as close o he eal sys em as possible.
Cu en applica ions use eal- ime ope a ing sys ems o ensu e
he high e esh a e needed o such simula ions and a e
usually qui e cos ly such as he dSPACE simula o o he
Na ional Ins umen s PXI sys ems jus o name a ew.
Real- ime compu ing is di ided in o so eal- ime, whe e
he use ulness o a esul deg ades a e i s deadline, i m eal-
ime whe e in equen deadline misses a e ole a ed and ha d
eal- ime whe e missing a deadline is o al sys em ailu e.
Missing deadlines deg ades he sys ems quali y he e o e a
ha d eal- ime simula ion me hod would gi e us he bes
esul s. The e a e nume ous eal- ime simula ions (RTS)
cu en ly in li e a u e [1-6], howe e hese sys ems a e so
eal- ime. Simila esea ch has been conduc ed by [7]
eaching simula ion s ep imes o 250-50 µs wi h a low cos
ha dwa e implemen a ion. In his pape expe imen s we e
concluded o see i a low cos implemen a ion using an ARM
p ocesso could achie e he same esul s o simple sys ems,
hus u he lowe ing he es ing cos o es ing, also p o iding
an oppo uni y o c ea e a sys em o he a e age use . Such a
sys em could also be used in educa ion due o he low p ice
and s ong connec ion o he cu en con ol sys ems cou se
ma e ial being augh a he Uni e si y o Szeged.
The desi ed end esul would be o c ea e a amewo k ha
enables ha d eal- ime ha dwa e in he loop es ing o linea
ime in a ian MIMO sys ems and sys ems wi h nonlinea
cha ac e is ics such as he pneuma ic a i icial muscle (PAM)
since he o he main esea ch opic a he Uni e si y o Szeged
Facul y o Enginee ing is he high accu acy posi ioning o
pneuma ic a i icial muscles o obo ics and ehabili a ion
applica ions [8].
II. METHODOLOGY
The hypo hesis being es ed was ha a 32bi ARM
mic ocon olle could be used o ha d eal- ime ha dwa e in
he loop simula ion. The mic ocon olle chosen was he
STM32F746ZGT6 on he STM32F746 Nucleo boa d (Fig. 1).
Fig. 1. STM32F746 Nucleo boa d
The STM32F746ZGT6 ea u es a 216 MHz ARM
p ocesso wi h a buil in single p ecision loa ing poin uni
enabling ope a ions on numbe s ep esen ed as loa s o be
comple ed unde a couple o cycles depending on he
ope a ion. A loa ing poin uni is o key impo ance as i
will be shown.
The p og amming en i onmen used was he Keil MDK
wi h he STM32CubeMX o pin se up and
ini ializa ion.Measu emen s we e done using LabVIEW, he
Recen Inno a ions in Mecha onics (RIiM) Vol. 3. (2016). No. 1-2.
DOI: 10.17667/ iim.2016.1-2/15.
da a acquisi ion de ice used was a myRIO (Fig. 2) om
Na ional Ins umen s.
Fig. 2. myRIO om Na ional Ins umen s
To es he hypo hesis a physical sys em was cons uc ed
consis ing o a small DC mo o wi h a magne moun ed on he
o a y sha and an AS5145 absolu e magne ic o a y encode
wi h 12 bi s o esolu ion moun ed pe pendicula o he sha
(Fig. 3).
Fig. 3. Sys em o be simula ed.DC mo o wi h magne and AS5145
magne icencode .
The p ocedu e used was he ollowing. Assuming he
model o he sys em o be he one shown on Fig. 4 whe e he
inpu o he sys em is ol age and he ou pu is he o a ional
speed o he sha equa ion (1) and (2) can be de i ed.
Fig. 4. Block diag am o mo o
whe e:
 – ol age di e ence be ween + and - inV
 i– cu en lowing h ough he mo o inA
 R – esis ance o he mo o inΩ
 L – induc ance o he mo o inH
 K – cons an o p opo ionali y inV/ ad/s
 J – ine ia inkgm2
 b – o a ional dampening cons an Nms/ ad

󰇗
– o o speedin ad/s
F om he block diag am equa ion (1) and (2) can be
de i ed.
(1)
(2)
Knowing equa ion (1) and (2) we can apply he Laplace
ans o m o hem. By exp essing he cu en we can subs i u e
(1) in (2) and exp ess he ans e unc ion o he mo o (3)
whe e he ol age is he inpu and o o speed is he ou pu .
( ) 󰇗
( )
( )
( ( ) )
(3)
Knowing he nomina o and denomina o o de s o he
ans e unc ion polynomials enables he pa ame ic sys em
iden i ica ion o he sys em once a s ep inpu o 1 V has been
applied and i s esponse measu ed. Since he o a y encode
only gi es ou absolu e posi ion he angula eloci y had o be
calcula ed. To calcula e he angula eloci y an FPGA
p og am had been w i en ha uns pe iodically e e y 400 µs.
The esul s ob ained and he esul o he sys em iden i ica ion
can be seen on Fig. 5.
Fig. 5. Resul s o measu emen (blue) and sys em iden i ica ion ( ed)
The ans e unc ion om sys em iden i ica ion can be
seen in equa ion (4)
Recen Inno a ions in Mecha onics (RIiM) Vol. 3. (2016). No. 1-2.
DOI: 10.17667/ iim.2016.1-2/15.
( ) 󰇗
( )
( )
( )
(4)
To implemen he acqui ed ans e unc ion on he
mic ocon olle a bilinea ans o ma ion was pe o med wi h
he sampling equency o 100kHz o ans o m i om he s
plane o he z plane gi ing us equa ion (5) wi h he ollowing
coe icien s.
( ) 󰇗
( )
( )
( )
(5)
Using an in e se z- ans o m he di e ence equa ion was
ob ained om he disc e e ime ans e unc ion. Implemen ing
he disc e e ime ans e unc ion is a simple ma e . Since he
equa ion only con ains addi ion mul iplica ion and sub ac ion
and each o hese ope a ions only equi es 1-3 cycles o he
ARMmic ocon olle .
III. RESULTS
Fig. 6 shows a measu emen o a s ep inpu applied o he
implemen ed ha dwa e in he loop sys em in con as wi h a
pu e ma hema ical model.
Fig. 6. Resul s o s ep inpu applied o ha dwa e in he loop (squa es) and
con inuous ime sys em model ( ed)
F om he compa ison i is isible ha he wo esponsesa e
close o each o he , wi h an a e age absolu e e o o 0.88004
ad/s we concluded ha using an ARMmic ocon olle is iable
o ha d eal- ime ha dwa e in heloop simula ion. The
STM32F746ZGT6 has a 12 bi analoginpu wi h a maximum
sampling equency o 2.4 MHz husplacing he maximum
sampling equency o he bilinea ans o ma ion a 2.4 MHz,
in iple in e lea e mode hesampling equency can be boos ed
o 7.2 MHz, doing so educes he capabili ies o he
mic ocon olle o a single inpu sys em.
REFERENCES
[1] S. Len ijo, A. Mon i, E. San i, C. Welch, and R. Dougal, “A new es ing
ool o powe elec onic digi al con ol,” in P oc. IEEE PESC, Jun.
2003, ol. 1, pp. 107–111.
[2] V. Dina ahi, R. I a ani, and R. Bone , “Design o a eal- ime digi al
simula o o a D-STATCOM sys em,” IEEE T ans. Ind. Elec on., ol.
51, no. 5, pp. 1001–1008, Jun. 2004.
[3] H. Li, M. S eu e , K. Shi, S. Wood u , and D. Zhang, “De elopmen o
a uni ied design, es , and esea ch pla o m o wind ene gy sys ems
based on ha dwa e-in- he-loop eal- ime simula ion,” IEEE T ans. Ind.
Elec on., ol.53,no.4,pp.1144–1151,Jun.2006.
[4] F. Piazza, S. Squa ini, R. Toppi, M. Na a i, M. Pon illo, F. Be a elli,
and A. La anzi, “Indus y-o ien ed so wa e-based sys em o quali y
e alua ion o ehicle audio en i onmen s,” IEEE T ans. Ind. Elec on.,
ol. 53, no. 3, pp. 855–866, Jun.2006.
[5] P. Ba acos, G. Mu e e, C. Rabba h, and W. Jin, “Enabling pc-based HIL
simula ion o au omo i e applica ions,” in P oc. IEEE IEMDC, Jun.
2001, pp. 721–729.
[6] M. Pa k and I. Yu, “A no el eal- ime simula ion echnique o
pho o ol aic gene a ion sys ems using RTDS,” IEEE T ans. Ene gy
Con e s., ol.19,no.1,pp.164–169,Ma .2004.
[7] Lu, Bin, Xin Wu, He nan Figue oa, and An onello Mon i. "A low-cos
eal- ime ha dwa e-in- he-loop es ing app oach o powe elec onics
con ols." IEEE T ansac ions on Indus ial Elec onics 54, no. 2 (2007):
919-931.
[8] Sá osi J.: “Accu a e Posi ioning o Humanoid Uppe A m” In e na ional
Jou nal o Enginee ing, Annals o Facul y o Enginee ing Hunedoa a,
2011, Vol. 9, No. Ex a, ISSN 1584-2673, pp. 33-36